Face age detection with CNN, RNN, transfer learning & scale-invariant feature transform (sift)

Automatic age estimation from real-world and unconstrained face images is rapidly gaining importance. In our proposed work, a deep CNN model that was trained on a database for face recognition task is used to estimate the age information on the Audience database. This paper has three significant contributions in this field. This work proves that a CNN model, which was trained for face recognition task, can be utilized for age estimation to improve performance; Over fitting problem can be overcome by employing a pre trained CNN on a large database for face recognition task; Not only the number of training images and the number subjects in a training database effect the performance of the age estimation model, but also the pre-training task of the employed CNN determines the model’s performance. Recently, many applications from biometrics, security control to entertainment use the information extracted from face images that contain information about age, gender, ethnic background, and emotional state. Automatic age estimation from facial images is one of the popular and challenging tasks that have different fields of applications such as controlling the content of the watched media depending on the customer's age estimation, whose objective is to determine the specific age or age group of a subject based on preliminary detected face region. Among its possible applications one should note electronic customer relationship management (such systems assume the usage of interactive electronic tools for automatic collection of age information of potential consumers in order to provide individual advertising and services to clients of various age groups), security control and surveillance monitoring The real-time audience measurement system consists of five consecutive stages: face detection, face tracking, gender recognition, age classification and in-cloud data statistics analysis. The challenging part of such system is age estimation algorithm on the basis of machine learning methods.

 The face aging process is determined by different factors: genetic, lifestyle, expression and environment. That is why same age people can have quite different rates of facial aging. We propose a novel algorithm consisting of two stages: adaptive feature extraction based on CNN classification. The feature extraction extracts feature corresponding to age and gender, while the classification classifies the face images to the correct age group and gender. Particularly, we address the large variations in the unfiltered real-world faces with a robust image pre-processing algorithm that prepares and processes those faces before being fed into the CNN model. Age and gender predictions of unfiltered real-life faces are yet to meet the requirements of commercial and real-world applications in spite of the progress computer vision community keeps making with the continuous improvement of the new techniques that improve the state of the art.

Keywords: Facial Age Image Dataset, CNN, RNN, Transfer learning methods (DenseNet, Resnet)